The proposed research, which is purely theoretical and computational, has three main parts, all focused on aspects of the phenomenon of molecular recognition, which is of basic and practical importance in biology and medicine. The first part concerns miniature receptor molecules, called hosts, which are used today to stabilize and deliver many drugs and which also show promise as medications in their own right. We plan to provide other scientists with new, tested software to help them design these miniature receptors and thereby speed the development of new medications. We also plan to carry out simulations of these molecules in order to develop a better understanding of how they work and also to gain insight into how larger receptors, such as many proteins, bind drugs. The second part is to study the changes in entropy that occur when molecules bind. In recent work, we have found that changes in entropy associated with the motions of receptors and the molecules they bind (ligands) can have a surprisingly strong influence on how tightly they bind each other. However, we do not yet understand these entropy changes well enough to make them work in our favor when designing tight-binding receptors and ligands. We plan to further develop our method of computing these entropy changes from computer simulations, and then use the method to develop a better understanding of them. For example, we would like to be able to predict when modifying a ligand to make it more rigid, and therefore lower in entropy, will increase its affinity. In addition, we plan to incorporate the new entropy calculations into software for computing binding affinities which we hope will help researchers design new drugs. The third part is to develop a new idea of applying the concept of stress to molecular biophysics. Materials scientists have come up with equations for computing the stress in a material from an atomistic computer simulation, and we think these equations can tell us something useful about how hosts, proteins and other molecules work. For one thing, we hypothesize that if receptor- ligand binding produces localized stress, then modifying the ligand to reduce this stress might increase the binding affinity. Thus, computing stress might help with the design of tight-binding ligands. We also hypothesize that, when a ligand binds an allosteric protein, a protein whose conformation shifts on binding, the mechanism of the shape change involves propagation of a wave of stress from the binding site. If we can understand how proteins change conformation, this would help us to re-engineer them for medical and industrial uses.